Have you really convinced yourself beyond any doubt that you need to be able to sense the identity of each piece as well as its presence?
I wanted to do this same thing. I looked at the DGT way of doing it and thought about RFID ideas, but in the end I used reed switches and magnets.
I followed the examples from my friend Berger here https://sites.google.com/site/bergersprojects/reedcb.
It works really well. I have it hooked up to a raspberry pi where the chessboard acts as a keyboard, and the display is a 4x20 LCD. I have a few buttons on the enclosure for interface. If you want to handle under promotion, you can use the buttons to indicate what piece to promote to. I run various UCI and winboard engines through polyglot http://wbec-ridderkerk.nl/html/details1/PolyGlot.html as a subprocess to my python script which handles all the interface.
It can also work with programs like Fritz or Arena on a PC or tablet.
I can also use it to just record PGN of live games against real people if I don't want to play against Stockfish or other cheaters like that.
Jimmy
It’s a nice work, Anyway why you gave up RFID solution?
I thought to design the antennas like this way.. Layer 02 And 03 is dual layer PCB
any comments ? :.
Chess Design.docx (83 KB)
kasun:
I thought to design the antennas like this way.. Layer 02 And 03 is dual layer PCB
What you have illustrated unfortunately makes no sense!
The function of a coil of wire - or a PCB track - is not dependent on whether you label it "Transmitter antenna" or "Receiver antenna" on a diagram, it is consequential to the shape and configuration of the wire.
In your diagram, current will preferentially flow through path "E" as it is the shortest route. The gauge of the wire or track has virtually no effect on the magnetic field though a wider track will have a lower resistance and tend to receive a larger share of the current in a parallel circuit. You also show "dead end" legs on each inner circuit where no current will flow at all, and the feeder lines on each side will tend to cancel the effect of the inner circuit anyway.
It may be that you anticipated making the inner loops of multiple "turns", in which case you are going to have to use feed-throughs into another layer to skip from the inside to the outside of the coil. This is in fact how printed RFID sense antennae are generally made.
In any case, what you need to do is to put the loops in series, and feed them from one side only. For single turn loops, this can be done by having each loop as a "U" and the return line down the edge of the board forms the fourth side. If using multi-turn loops, then the start and end of the loop will be on the same side.
Having a complete "donut" of copper in another plane will completely block any effect of both transmitting and receiving loops. Conversely, having an incomplete ("C") loop will have no effect - you could however use a relatively broad area (separately) between each of the individual loops to constrain the field.
Grumpy_Mike:
Are you saying though that polymorphs idea of having eight transmitters for each row and eight receivers for each column would not work?
Yes I am. The problem is that there will be inductive coupling between the antenna so no amount of switching is going to isolate one token.
I can see that you will get a certain amount of inductive coupling all the time, but I'm wondering whether it can be made to work with a few modifications:
-
Use a small mosfet instead of a triac to short out the receive coils that are not in use. The mosfet will short them more effectively, although it will also mean that the received voltage has to be less than about 0.5V peak by design, so that the body diode doesn't conduct.
-
Instead of looking for an on/off signal, measure the amplitude of the signal using a phase sensitive detector. The wanted signal should be phase-shifted from the unwanted signal, so the phase of the reference signal can be adjusted to make the detector less sensitive to the unwanted signal. Then use a comparator (perhaps the mcu built-in comparator) to set an amplitude threshold.
Have you really convinced yourself beyond any doubt that you need to be able to sense the identity of each piece as well as its presence?
You don't need identity but pawn promotion poses a problem you cannot solve (as far as I can see) without also keeping track of major pieces (i.e. not pawns) that have been captured so you know what piece a pawn has been promoted to.
In other words you need to be able to figure out what piece is on each of the 64 squares but also have an addition 14 squares off-board to hold captured pieces.
The RFID solution identifies pieces and so is elegant, but the electronics seem to be getting more complicated.
Mixographer you used reed switches, how did you handle pawn promotion, did you ignore it, or assume pawns always became queens, or do you provide additional data to the system manually?
A resistor on the base of the piece.
On a 8x8 matrix.
- A B C D E F G H
1
2
3
4
5
6
7
8
Scan each matrice for changes in resistance, so if a pawn from B2 was picked up, the resistance in column B & row 2 changes, and were the piece lands the resistance changes. i.e. B4
Because the board is always set the same way, and white always goes first. You could also track which piece was taken, and which taken piece is replaced.
Looks fun, haven't played in over a decade 
How do you account for dirty contacts or incomplete connection?
How do you account for dirty contacts or incomplete connection?
You may have well asked, how do you stop magnetic interference?
You may have well asked, how do you stop magnetic interference?
Not really. Having electrical contacts all over the board sounds messy and will be distracting visually as well as because of the need to place pieces carefully. Magnetic fields are not an issue in any of the chess settings I can think of.
Magnetic fields are not an issue in any of the chess settings I can think of
How does rfid work ?
JB_AU:
Magnetic fields are not an issue in any of the chess settings I can think of
How does rfid work ?
See my reply #44 for an approach that might be sufficient to handle the unwanted interference from other pieces above the same transmitting coil.
JB_AU:
How do you account for dirty contacts or incomplete connection?
You may have well asked, how do you stop magnetic interference?
What does that have to do with dirty contacts or incomplete connection? You don't prove your idea by disproving some other idea.
I believe I and others have already talked about how to make an RFID system work.
In reply to Radman:
For pawn promotion, the idea is that the board always provides two step moves like e2e4 or e7e5. When a pawn moves to the eigth rank, you send an additional keystroke to represent the promoted piece. So it becomes like h7h8R (Rook).
Jimmy
Is there no way round this by say reducing the transmitter power (the rfid tags will be right on top of the transmitters) and by grounding the transmitters that are not supposed to fire?
No it is not a matter of reducing the power the inductive coupling is such that that whole lot in effect shorts out. I realised that doing the RFID sequencer project.
Sent from an Irish bar in the middle of Rome watching Scotish football - go figure.
Looks fun, haven't played in over a decade
If you visit this excellent site you will get a game, for free, in a few seconds (I suggest unticking 3mins and 5mins and ticking 10mins).
The solution using magnets in pieces to activate reed switches is not going to be affected by stray magnetic fields.
The solution of having resistors in pieces will be very dependent on making good contacts.
Magnets are a better solution. The RFIDs in theory are better still - if there is a practical way to implement them, however if Grumpy-Mike thinks there is an issue that needs to be seriously considered.
Look at the situation attached.
One excitation coil and three pick up coils with one token.
All the pick up coils pick up the same signal, not just the one over the token. This is because the signal generated is received in pickup coil one, but the excitation coil couples all the coils together so all the pickup coils produce exactly the same signal.
RFID coils.pdf (14.1 KB)
Oh, fer cryin' out lawd... detune the coils not in use, don't short them out.
I came here to offer an idea, i didn't come to offer a final solution.
Every idea will have pros & cons, for every idea.
If its expected that every answer needs to be scrutinized, to be a valid response, then this forum which shares ideas, creativity & community interaction ceases to be a place of openness.
